ИИ от А до Я вся теория. LLM, токены и вероятности, Нейронные сети, RAG, MCP, Агенты и Harness
Ulbi TV
294,093 views
In this video, we'll cover the basics of AI in one lesson. We'll cover neural networks, LLM, tokens and vectors, probability distributions, the Transformer architecture, and the attention mechanism. We'll also cover context and context windows, RAG, tool calling, and MCP. We'll also compare vibe coding and AI engineering.
Vibe coding roadmap - https://itbooster.ru/roadmap/28/
"Techies" community + architectural content - https://boosty.to/ulbitv
My Telegram channel - https://t.me/ulbi_tv
IT job openings - https://t.me/+Cf8ezLjjm8JmNzky
Timecodes:
00:00 ➝ Introduction
00:30 ➝ Lesson plan
01:20 ➝ Introduction to the theory. Algorithms and AI. Neural networks and LLM. Training
04:40 ➝ LLM as a Pure Function. Tokens, Probability Distributions, and Temperature. Vectors (Embeddings)
09:00 ➝ Context and Context Window. Overflow and Optimization.
11:20 ➝ The Meaning of Words. Transformer Architecture and Attention. The Beginning of GPT
13:00 ➝ RAG. Chunks, Vectors, and Vector Database
15:20 ➝ Hallucinations
16:25 ➝ Tools. Tool Calling
17:40 ➝ External Services and MCP
18:50 ➝ Agents and Their Architecture. Agent Loop
20:20 ➝ Agent Shells, Harness. Agent Orchestration
22:15 ➝ Skills
23:00 ➝ Model Selection
23:30 ➝ Vibe Coding vs. AI Engineering AI Expertise
27:00 ➝ Thank you all for your support. If you're interested, don't forget to like and comment.
Link to my Telegram channel - https://t.me/ulbi_tv
You can support me and my channel using the links below.
"Techies" Community + additional content - https://boosty.to/ulbitv
20 Comments
ИИ от А до Я вся теория. LLM, токены и вероятности, Нейронные сети, RAG, MCP, Агенты и Harness
Ulbi TV
294,093 views
In this video, we'll cover the basics of AI in one lesson. We'll cover neural networks, LLM, tokens and vectors, probability distributions, the Transformer architecture, and the attention mechanism. We'll also cover context and context windows, RAG, tool calling, and MCP. We'll also compare vibe coding and AI engineering.
Vibe coding roadmap - https://itbooster.ru/roadmap/28/
"Techies" community + architectural content - https://boosty.to/ulbitv
My Telegram channel - https://t.me/ulbi_tv
IT job openings - https://t.me/+Cf8ezLjjm8JmNzky
Timecodes:
00:00 ➝ Introduction
00:30 ➝ Lesson plan
01:20 ➝ Introduction to the theory. Algorithms and AI. Neural networks and LLM. Training
04:40 ➝ LLM as a Pure Function. Tokens, Probability Distributions, and Temperature. Vectors (Embeddings)
09:00 ➝ Context and Context Window. Overflow and Optimization.
11:20 ➝ The Meaning of Words. Transformer Architecture and Attention. The Beginning of GPT
13:00 ➝ RAG. Chunks, Vectors, and Vector Database
15:20 ➝ Hallucinations
16:25 ➝ Tools. Tool Calling
17:40 ➝ External Services and MCP
18:50 ➝ Agents and Their Architecture. Agent Loop
20:20 ➝ Agent Shells, Harness. Agent Orchestration
22:15 ➝ Skills
23:00 ➝ Model Selection
23:30 ➝ Vibe Coding vs. AI Engineering AI Expertise
27:00 ➝ Thank you all for your support. If you're interested, don't forget to like and comment.
Link to my Telegram channel - https://t.me/ulbi_tv
You can support me and my channel using the links below.
"Techies" Community + additional content - https://boosty.to/ulbitv